Radiosynthesis and characterization of astemizole derivatives as lead compounds toward PET imaging of τ-pathology

Article abstract of DOI:10.1039/c3md00017f

Formation of neurofibrillary tangles, comprising of microtubule-associated tau protein, is a hallmark of a group of neurodegenerative diseases, including Alzheimer's disease. In consequence, in vivo imaging of neurofibrillary tangles is a current focus of positron emission tomography research. Herein, development of an in vitro radioligand binding assay which uses synthetic aggregates as a model of neurofibrillary tangles is reported, together with evaluation of novel derivatives of the tau protein ligand astemizole. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3md00017f

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Radiosynthesis and characterization of astemizole
derivatives as lead compounds toward PET imaging
of s-pathology†
Cite this: Med. Chem. Commun., 2013,
4, 852
abc
Laurent Brichard,a Valentina Ferrari,a David J. Williamson,a
*
Patrick J. Riss,
Tim D. Fryer,ab Young T. Hong,a Jean-Claude Barona and Franklin I. Aigbirhioab
Formation of neurofibrillary tangles, comprising of microtubule-associated tau protein, is a hallmark of a
group of neurodegenerative diseases, including Alzheimer's disease. In consequence, in vivo imaging of
Received 14th January 2013
Accepted 13th March 2013
neurofibrillary tangles is
a current focus of positron emission tomography research. Herein,
development of an in vitro radioligand binding assay which uses synthetic aggregates as a model of
neurofibrillary tangles is reported, together with evaluation of novel derivatives of the tau protein
ligand astemizole.
DOI: 10.1039/c3md00017f
www.rsc.org/medchemcomm
Dementia is characterised by progressive loss of memory and candidates, [18F]FDDNP, suffers from low specicity for tau as it
cognitive impairment, ultimately resulting in functional is a dual Ab-tau ligand.5 More recent reports outline develop-
disability.1 These symptoms are caused by neurodegeneration ment and preclinical characterization of more selective NFT
in the brain. Various forms of dementia can be distinguished radioligands, which might be suitable candidates for PET
based on characteristics of the underlying molecular pathology. imaging.6 Rojo et al. discovered that the commercially available
Oen, post-mortem detection of pathological protein deposits compounds lansoprazole and astemizole bind to NFT derived
formed from degraded proteins, such as b-amyloid (Ab), from human brain tissue, as well as to synthetic tau brils
microtubule associated tau protein (s-protein) and a-synuclein, composed of recombinant human tau-441.7
A carbon-11
is required for a denitive diagnosis and classication of labelled derivative of lansoprazole was reported, showing
dementia.2 The gold standard for the diagnosis of Alzheimer's selective accumulation in NFT-rich regions of mouse brains.8
disease (AD) is the presence of both intracellular neurobrillary
We have focused our effort on astemizole, to develop a
tangles (NFT), comprising of aggregated paired helical la- radioligand labelled with 18F for PET imaging of NFT burden in
ments of hyper-phosphorylated s-protein, and extracellular Ab- animal models and human subjects. In order to determine the
plaques consisting of brillary Ab aggregates. Positron emission binding affinity of novel derivatives, we developed a radioligand
tomography (PET) imaging may provide an in vivo alternative to binding assay based on previously reported procedures.7,9
post-mortem diagnosis of protein aggregates, potentially in pre- Availability of this straightforward assay will simplify structure–
symptomatic stages, because NFT formation occurs before activity relationship studies for the development of NFT ligands
clinical symptoms become apparent and the NFT burden by methodological modication of known lead structures. The
correlates with the severity of dementia symptoms.3 Hence, the use of such assays as reliable in vitro methods for determination
development of a selective PET radiotracer for non-invasive of binding affinity is well established. Contrary to uorescent
quantitative imaging of NFT in vivo would reinforce insights probes, radioactivity is insensitive to concentration effects and
into the aetiology of s-pathologies, and facilitate monitoring of fully quantiable even at nanomolar concentrations. Synthetic
therapy effect and efficiency.4
aggregates formed from recombinant s-isoforms provide an
So far, PET imaging of NFT in dementia patients has been inexpensive model of native NFT.10 Thus, inhibition of radio-
impeded by the lack of suitable radioligands. One of the earliest ligand binding to synthetic aggregates provides a cost effective
method for screening novel small molecules.
An automated procedure for the radiosynthesis of [11C]
aWolfson Brain Imaging Centre, University of Cambridge, Addenbrooke's Hospital, CB2
0QQ, Cambridge, UK. E-mail: pr340@wbic.cam.ac.uk; Fax: +44 (0)1223 331826; Tel:
astemizole was developed, which was then used as the tau-
binding radioligand, analogous to the work of Rojo et al.
Contrary to Rojo et al., we resorted to small pore-size nitrocel-
+44 (0)1223 331823
bBehavioural and Clinical Neuroscience Institute, University of Cambridge,
Cambridge, UK
lulose membranes for ltration, which had been found to be
cDepartment of Chemistry, Universitetet i Oslo, Postboks 1033, Blindern, Oslo, Norge.
advantageous for tau immuno-staining.9 Our initial experi-
E-mail: Patrick.Riss@kjemi.uio.no; Tel: +47 228 57673
ments were conducted using a cell harvester, glass tubes and
† Electronic supplementary information (ESI) available: Experimental procedures.
glass bre lters with the aim of reproducing the reported
See DOI: 10.1039/c3md00017f
852 | Med. Chem. Commun., 2013, 4, 852–855
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
Table 1 Radioactive and non-radioactive astemizole derivatives
value of 2.6 Æ 0.4 nM, indicating good congruence of our results
with previously published values, within the error margin of
pipetting and liquid scintillation counting. Competitive
binding of [11C]astemizole with lansoprazole was used as a
control experiment. Ki values were computed using a one
binding site model as implemented in Graphpad Prism 6 so-
ware. Lansoprazole was found to have an inhibition potency of
13.9 nM in our assay, in compliance with the ndings reported
by Rojo et al., which highlights the accuracy of the employed
methodology. The validated assay was successfully transferred
to 350 mL 96-well lter plates to facilitate parallel screening of
multiple compounds per experiment. In order to optimise
throughput, [3H]astemizole was used in all further experiments.
The solubility of astemizole in buffer is very limited and non-
specic binding or precipitation might confound the experi-
mental results. We resorted to using 8% aqueous ethanol as
assay medium and reduced the maximum concentration of the
ligand to 5 mM. The suitability of this assay for assessment of
NFT affinity was investigated using known s-binding
compounds in order to provide a comparative benchmark. We
found Ki values for our controls lansoprazole, methylene blue
and thioavin S in accordance with literature values (Table 1).
Aer successful validation, the novel assay was used in a
proof of principle study using four structural analogues of
astemizole, designed with the aim of incorporating a second
Radiochemical
yield/%
Compound
Ki/nMa
Yield/%
1
2
4a
4b
4c
4d
4e
5
2.6 Æ 0.4b
7.0 Æ 0.4
1.67 Æ 0.38
1.93 Æ 0.36
3.36 Æ 0.58
5.47 Æ 0.41
4.2 Æ 0.9
—
—
66
89
90
86
82
69
77
—
—
—
2–5c
—
29d, 5–10e
—
—
—
24e
—
—
—
—
Lansoprazole
Thiazine acetate
Thioavin S
13.9 Æ 1.0
78.5
>1000
a
b
The concentration for [3H]astemizole in the assay was 0.5 Â kd. kd
value. Labelled with 11C. Directly labelled with 18F. Indirectly
labelled with 18F. Radiochemical yields are non-decay corrected yields
relative to end-of-bombardment.
c
d
e
ndings. [11C]Astemizole was obtained via captive solvent
methylation of compound 2 using [11C]CH3I in the presence of
NaOH in DMF. The kd value for [11C]astemizole was determined
and compared to the literature value.8 Scatchard analysis of a
saturation binding experiment on synthetic brils obtained
from heparin-induced brilisation of htau-441 yielded a kd
Scheme 1 Synthetic pathways to compounds 2–5.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 852–855 | 853

View Article Online
MedChemComm
Concise Article
uorine atom bound to an aliphatic carbon atom, allowing for
direct nucleophilic radiouorination. To achieve this, astemi-
zole (1) was demethylated in reuxing 48% aqueous HBr solu-
tion overnight and the product was isolated by fractional
precipitation in a mixture of aqueous ammonia and DMF to
obtain (2) with a 66% yield.11 Compound 2 was alkylated with
ꢀ
3a–e using potassium carbonate and acetone at 70 C (Scheme
1; Table 1) to yield the desired uoroalkyl derivatives 4a–e.
Determination of Ki values for compounds 2 and 4a–e showed
that the 4-position of the phenylethyl moity in astemizole is
relatively tolerant towards modication (Table 1). The short 2-
uoroethyl and 3-uoropropyl chain ethers in compounds 4a
and 4b, respectively, were found to be well tolerated, with no
pronounced inuence on inhibition potency. Derivatives 4c–d
showed higher Ki values which nonetheless were within the
single gure nanomolar range. Phenol 2 showed approximately
fourfold lower binding affinity than 4a.
Fig. 1 SUV maps of [18F]4a 120–150 minutes post-injection under cyclosporine A
pretreatment (left) and in an untreated animal (right), overlaid on the brain MRI
template. White outline denotes cortical ROI, green outlines denote left and right
striatal ROI, blue outlines denote hippocampal ROI, yellow outline denotes hypo-
thalamic ROI.
In order to investigate a radiolabelled analogue of 4a, alcohol
4d was converted into p-toluenesulfonyl ester 5 using p-tolue-
nesulfonyl chloride and triethyl amine in dichloromethane at
0
ꢀC with a yield of 77% and used as labelling precursor for
direct nucleophilic radiouorination. Via direct nucleophilic
substitution in MeCN under reux, radiochemical yields of up
to 29% were achieved. In addition, compounds [18F]4a and [18F]
4e were synthesised from phenol 2 using two step procedures
via synthesis of the 18F-labelled secondary labelling reagents
[18F]3a and [18F]3e,12 followed by O-alkylation. In principle, this
indirect route can produce the novel candidates 4a–e without
the need for individual labelling precursors. We focused our
effort on the development of a universal 1-pot method for the
automated production of [18F]4a using a cGMP compliant GE
TRACERlab FX F-N radiosynthesis module. In the rst part of
this method, [18F]uoride was allowed to react at 90 ꢀC with
ethylene ditosylate in acetonitrile to form [18F]3a. Intermediate
[18F]3a was subsequently subjected to reaction with the phenol
2 and a base in DMF at 120 ꢀC to give the nal product [18F]4a.
To investigate the feasibility of [18F]4a imaging in rat models
of tau pathology, a pre-clinical PET study was conducted in
healthy male Wistar rats using a microPET Focus 220. It has
been surmised that astemizole is a substrate of the ATP-
dependent efflux transporter p-glycoprotein (p-gp) in rats, which
would inuence its utility for PET.13 To assess the effect of p-gp
inhibition, one rat was pretreated with the p-gp inhibitor
cyclosporine A (50 mg kgÀ1). Animals were imaged for 2.5 hours
post [18F]4a injection, then immediately sacriced. Blood,
Fig. 2 Regional time activity curves of [18F]4a in the cyclosporine A pretreatment
rat (solid lines) and the untreated rat (dashed lines).
Table 2 In vivo evaluation of [18F]4a
Cyclosporine A
treated animal
Untreated animal
Injected activity
Body weight
38.2 MBq
579 g
13.5
26.4 MBq
550 g
1.62
SUV in blooda
SUV in plasmaa
5.98
0.78
plasma and brain tissue were analysed for radiolabelled % Intact in plasmaa
53.4 Æ 0.2
44.3 Æ 2.1
84%
0.46
0.27
54 Æ 1
48.5 Æ 2.4
96%
0.19
0.10
% Unbound in plasmaa
metabolites using radioTLC and radioHPLC. Radioactivity
concentrations in blood and plasma and the radioligand free
% Intact in braina
SUV in frontal cortexb
fraction were determined. Images reconstructed with ltered
back projection were co-registered to a rat brain MRI template
and converted to standardized uptake value (SUV; Fig. 1).14
Regions of interest drawn on the brain template were used to
produce regional time–activity curves (TACs, Fig. 2) and SUV
values (Table 2).
SUV in striatumb
SUV in hippocampusb
SUV in hypothalamusb
0.31
0.32
0.12
0.10
a
150 min post injection. b 120–150 min post-injection.
Unfortunately, compound [18F]4a showed low brain uptake, aer an initial peak, whereas with blockade the TACs became
with low SUV values found even aer saturation of p-gp. virtually constant over time. A signicant contribution to the
Without blockade, the TACs indicated washout from the brain TACs will be signal from the vascular compartment (ꢁ4% of
854 | Med. Chem. Commun., 2013, 4, 852–855
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
brain tissue), especially for the pretreated animal whose blood
radioactivity concentration was 8 times higher 150 min post-
injection, possibly due to cyclosporine A reducing renal excre-
tion. [18F]4a was metabolised to a single polar metabolite, which
accounted for approximately 46% of the plasma radioactivity
concentration 150 min post-injection. Cyclosporine A pretreat-
ment increased the metabolite fraction found in brain.
J. Arteaga, D. K. Cashion, G. Chen, U. Gangadharmath,
L. F. Gomez, D. Kasi, C. Lam, Q. Liang, C. Liu,
V. P. Mocharla, F. Mu, A. Sinha, A. K. Szardenings,
E. Wang, J. C. Walsh, C. Xia, C. Yu, T. Zhao and
H. C. Kolb, J. Alzheimer's Dis., 2012, 31, 601; (c)
M. T. Fodero-Tavoletti, N. Okamura, S. Furumoto,
R. S. Mulligan, A. R. Connor, C. A. McLean, D. Cao,
A. Rigopoulos, G. A. Cartwright, G. O'Keefe, S. Gong,
P. A. Adlard, K. J. Barnham, C. C. Rowe, C. L. Masters,
Y. Kudo, R. Cappai, K. Yanai and V. L. Villemagne, Brain,
2011, 134, 1089; (d) N. Okamura, T. Suemoto, S. Furumoto,
M. Suzuki, H. Shimadzu, H. Akatsu, T. Yamamoto,
H. Fujiwara, M. Nemoto, M. Maruyama, H. Arai, K. Yanai,
T. Sawada and Y. Kudo, J. Neurosci., 2005, 25, 10857.
7 L. E. Rojo, J. Alzate-Morales, I. N. Saavedra, P. Davies and
R. B. Maccioni, J. Alzheimer's Dis., 2010, 19, 573.
Conclusions
Compounds 4a–e were developed and evaluated as potential
radiotracers for tau-pathology. [18F]4a shows strong binding
affinity to synthetic tau brils but in vivo imaging revealed low
brain uptake, even with efflux transporter inhibition, which will
limit the utility of this tracer. However, compounds 1 and 4a–e
are still useful as radioligands for in vitro studies of tau-
pathology and in particular screening of tau binding affinity. In
addition, these compounds provide a key lead structure for the
development of second generation derivatives with better
solubility and brain uptake.
8 X. Shao, G. M. Carpenter, T. J. Desmond, P. Sherman,
C. A. Quesada, M. Fawaz, A. F. Brooks, M. R. Kilbourn,
R. L. Albin, K. A. Frey and P. J. H. Scott, ACS Med. Chem.
Lett., 2012, 3, 936.
9 E. Chang and J. Kuret, Anal. Biochem., 2008, 373, 330.
10 (a) D. M. Wilson and L. I. Binder, Am. J. Pathol., 1997, 150,
2181; (b) S. Barghorn, P. Davies and E. Mandelkow,
Biochemistry, 2004, 43, 1694; (c) C. N. Chirita,
E. E. Congdon, H. Yin and J. Kuret, Biochemistry, 2005, 44,
5862.
11 F. Janssens, J. Torremans, M. Janssen, R. A. Stokbroekx,
M. Luyckx and P. A. J. Janssen, J. Med. Chem., 1985, 28,
1934.
12 (a) P. J. Riss and F. I. Aigbirhio, Chem. Commun., 2011, 47,
11873; (b) P. J. Riss, V. Ferrari, L. Brichard, P. Burke,
R. Smith and F. I. Aigbirhio, Org. Biomol. Chem., 2012, 10,
6980.
Notes and references
1 M. Goedert and M. G. Spillantini, Science, 2006, 314, 777.
2 P. Friedhoff, M. von Bergen, E. M. Mandelkow and
E. Mandelkow, Biochim. Biophys. Acta, 2000, 1502, 122.
3 (a) S. Kim, J. R. Jensen, K. Cisek, K. E. Funk, S. Naphade,
K. Schafer and J. Kuret, Curr. Alzheimer Res., 2010, 7, 230;
(b) M. Goedert, A. Klug and R. A. Crowther, J. Alzheimer's
Dis., 2006, 9(suppl. 3), 195.
4 (a) C. A. Mathis, Y. Wang and W. E. Klunk, Curr. Pharm. Des.,
2004, 10, 1469; (b) D. R. Williams, Intern. Med. J., 2006, 36,
652.
13 L. Cai, J. S. Liow, B. Houlihan, C. Morse, R. B. Innis and
V. W. Pike, NeuroImage, 2010, 52, S152.
5 (a) E. D. Agdeppa, V. Kepe, J. Liu, S. Flores-Torres,
N. Satyamurthy, A. Petric, G. M. Cole, G. W. Small,
S.-C. Huang and J. R. Barrio, J. Neurosci., 2001, 21, 1; (b)
P. W. Thompson, L. Ye, J. L. Morgenstern, L. Sue,
T. G. Beach, D. J. Judd, N. J. Shipley, V. Libri and
A. Lockhart, J. Neurochem., 2009, 109, 623.
6 (a) K. Matsumura, M. Ono, S. Hayashi, H. Kimura,
Y. Okamoto, M. Ihara, T. Ryosuke, M. Hiroshi and
S. Hideo, Med. Chem. Commun., 2011, 2, 596; (b) W. Zhang,
14 (a) P. A. Valdes-Hernandez, A. Sumiyoshi, H. Nonaka,
R. Haga, E. Aubert-Vasquez, T. Ogawa, Y. Iturria-Medina,
J. J. Riera and R. Kawashima, Front. Neuroinform., 2011, 5,
26; (b) P. J. Riss, Y. T. Hong, D. Williamson, D. Caprioli,
S. Sitnikov, V. Ferrari, S. J. Sawiak, J. C. Baron, J. W. Dalley,
T. D. Fryer and F. I. Aigbirhio, J. Cereb. Blood Flow Metab.,
2011, 31, 2334.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 852–855 | 855